Amygdala damage affects event-related potentials for fearful faces at specific time windows

r Human Brain Mapping 31:1089–1105 (2010) r

Amygdala Damage Affects Event-RelatedPotentials for Fearful Faces at Specific

Time Windows

Pia Rotshtein,1,2* Mark P. Richardson,3 Joel S. Winston,2 Stefan J. Kiebel,2,4

Patrik Vuilleumier,5 Martin Eimer,6 Jon Driver,2,7 and Raymond J. Dolan2

1School of Psychology, University of Birmingham, Edgbaston, Birmingham, United Kingdom2Wellcome Centre for Neuroimaging at UCL, University College London, London, United Kingdom

3Department of Clinical Neuroscience, King’s College London Institute of Psychiatry, London,United Kingdom

4Department of Neurology, Max Planck Institute for Human Cognitive and Brain Sciences,Leipzig, Germany

5Laboratory for Neurology and Imaging of Cognition, Department of Neuroscience and Clinic ofNeurology, University of Geneva Medical Centre, Geneva, Switzerland

6School of Psychology, Birkbeck College, London, United Kingdom7UCL Institute of Cognitive Neuroscience, University College London, London, United Kingdom

r r

Abstract: The amygdala is known to influence processing of threat-related stimuli in distant brain regions,including visual cortex. The time-course of these distant influences is unknown, although this informationis important for resolving debates over likely pathways mediating an apparent rapidity in emotional proc-essing. To address this, we recorded event-related potentials (ERPs) to seen fearful face expressions, in pre-operative patients with medial temporal lobe epilepsy who had varying degrees of amygdala pathology,plus healthy volunteers. We found that amygdala damage diminished ERPs for fearful versus neutral faceswithin the P1 time-range, �100–150 ms, and for a later component at �500–600 ms. Individual severity ofamygdala damage determined the magnitude of both these effects, consistent with a causal amygdala role.By contrast, amygdala damage did not affect explicit perception of fearful expressions nor a distinct emo-tional ERP effect at 150–250 ms. These results demonstrate two distinct time-points at which the amygdalainfluences fear processing. The data also demonstrate that while not all aspects of expression processing aredisrupted by amygdala damage, there is a crucial impact on an early P1 component. These findings are con-sistent with the existence of multiple processing stages or routes for fearful faces that vary in their depend-ence on amygdala function. Hum Brain Mapp 31:1089–1105, 2010. VC 2009Wiley-Liss, Inc.

Keywords: ERP; medial temporal lobe epilepsy; emotion; P1; late-P3; SPM5

r r

Additional Supporting Information may be found in the onlineversion of this article.

Contract grant sponsors: Wellcome Trust Programme Grant (to RJD,JD), Wellcome Trust funding (to PR, and SK), MRC (JD and ClinicianScientist Fellowship to MPR), ESRC-MRC (Fellowship to PR); SwissNational Science Foundation (to PV), Royal Society (to ME and JD).

*Correspondence to: Pia Rotshtein, School of Psychology, Univer-sity of Birmingham, Edgbaston, Birmingham B15 2TT, UK.

E-mail: P. [email protected]

Received for publication 17 June 2009; Revised 26 August 2009;Accepted 9 September 2009

DOI: 10.1002/hbm.20921Published online 16 December 2009 in Wiley InterScience (www.interscience.wiley.com).

VC 2009 Wiley-Liss, Inc.

INTRODUCTION

The amygdala is considered pivotal to processing emo-tional information [Dolan, 2002; Phelps and LeDoux, 2005],notably for threat-related stimuli [Amaral, 2002]. Projec-tions from the amygdala are hypothesized to orchestrateadaptive behavioral and autonomic reactions [Amaral,2002; LeDoux, 1996], modulate emotional face processingin extrastriate visual cortex [Morris et al., 1998; Noesseltet al., 2005; Rotshtein et al., 2001; Vuilleumier et al., 2004],and influence encoding of emotional stimuli into long-term memory [Dolan, 2002; Richardson et al., 2004; Smithet al., 2006]. Amygdala lesions abolish differential BOLDresponses to fearful stimuli (vs. neutral) in extrastriateregions [Vuilleumier et al., 2004]. This latter observationaccords with a causal role for the amygdala in mediatingmodulatory influences on sensory cortices [Amaral, 2003;Armony et al., 1998; LeDoux, 1996; Phelps and LeDoux,2005]. However, the sluggish nature of the BOLD responsemeans that fMRI data alone cannot reveal the time-courseof remote amygdala influences. Such information is impor-tant to establish whether amygdala emotional responsesprecede or follow other perceptual and face-selectiveresponses, such as those within visual cortex (see detaileddiscussion below).

Evoked related potential (ERP) studies in healthy volun-teers have reported effects of threat-related stimuli (e.g.seen fearful faces) at several time windows. The earliestdifferential response for fearful versus neutral faces isfound around �120 ms poststimulus onset [Eger et al.,2003; Eimer and Holmes, 2002; Pourtois et al., 2004, 2005a;Streit et al., 2003; van Heijnsbergen et al., 2007], corre-sponding to the visual P1 component that has been arguedto reflect perceptual processing [Aru and Bachmann, 2009;Marzi et al., 2000; Thorpe et al., 1996; though see Brissonand Jolicoeur [2008]]. The P1 component is hypothesizedto be generated in posterior occipito-temporal areas [DiRusso et al., 2002]. Emotional effects in this time windowhave also been reported for subliminal stimuli [Liddellet al., 2004]. The time window of this early scalp ERPeffect is in line with initial response latencies of the amyg-dala shown by intracranial recording in both humans andmonkeys [Gothard et al., 2007; Krolak-Salmon et al., 2004;Oya et al., 2002], though may have a different source. Asecond component, at around �200 ms for fronto-centralelectrodes, shows differential responses to threat stimulidepending on the intensity on the depicted emotion andits perception [Ashley et al., 2004; Eimer et al., 2008;Holmes et al., 2003; Leppanen et al., 2007; Liddell et al.,2005; Sprengelmeyer and Jentzsch, 2006]. This subsequentERP effect may last up to 1 s poststimulus presentation[Eimer and Holmes, 2007]. Emotional modulations are alsobeen reported in the human amygdala within this timerange [Krolak-Salmon et al., 2004]. Finally, a late effect ofemotions on ERPs is usually seen over central electrodesafter �500 ms [Johansson et al., 2004; Keil et al., 2001; Kiss-ler et al., 2006] and is associated with influences of emo-

tion on episodic memory processes [Liddell et al., 2004;Maratos et al., 2000]. It remains entirely unknown whetherprojections from the amygdala contribute to these differentERP effects measured at the scalp. Specifically, it is unre-solved how early in time amygdala influences on distantcortical regions can arise in humans and which (if any)threat-related ERP components reflect direct influencesfrom the amygdala.

Two opposing perspectives have been advanced regard-ing the possible time-course of amygdala involvement inprocessing threat-related information. The first emphasizesthat the amygdala receives information following process-ing by the corresponding sensory cortices and possibly af-ter conscious perception by the observer [Pessoa et al.,2005a,b]. In accord with this, some neuroimaging studiesin healthy humans have indicated that some aspects of theBOLD response in the amygdala depend on conscious per-ception of the stimulus and/or attention [Holmes et al.,2003; Pessoa et al., 2002, 2005a; Phillips et al., 2004].According to this perspective, one might anticipate thatamygdala function should influence threat-related ERPsrelatively late in poststimulus onset, e.g., perhaps after�150 ms when the major effects of attention are typicallyobserved [Di Russo et al., 2003] and when perception maybecome conscious [Thorpe et al., 1996], including for faceexpressions [Ashley et al., 2004; Eimer et al., 2008; Holmeset al., 2003; Leppanen et al., 2007; Liddell et al., 2005;Sprengelmeyer and Jentzsch, 2006].

An alternative perspective suggests that a rapidresponse, initiated by the amygdala, allows a fast modula-tion of sensory processing by the amygdala, especially forinformation containing threat-related signals [Compton,2003; Dolan and Vuilleumier, 2003; LeDoux, 1996; Ohmanet al., 2007; Ohman, 2005; Phelps and LeDoux, 2005; Vuil-leumier, 2005]. This perspective receives support fromneuropsychological studies of ‘‘blindsight’’ patients withlesions to striate and extrastriate cortex, who still retainsome ability to respond to the presence of threat-relatedstimuli [de Gelder et al., 1999; Pegna et al., 2005] and canshow preserved amygdala response during fMRI [deGelder et al., 2005; Morris et al., 2001; Pegna et al., 2005].Neuroimaging studies in healthy participants have alsoshown amygdala responses to subliminal threat stimuli[Carlsson et al., 2004; Morris et al., 1999] and to stimulioutside the main focus of attention [Vuilleumier et al.,2001]. From this perspective, one might anticipate thatamygdala influences upon visual ERPs might emerge rela-tively early, possibly even for the first ERP effects pro-duced by emotion expressions, i.e., for the P1 componentat �120 ms. However, the actual time-course of amygdalacontributions to threat-related ERPs has never beendirectly investigated in humans previously.

A further controversy is whether intact amygdala func-tion is necessary for explicit perception of seen emotionalexpressions, particularly fear [Cristinzio et al., 2007;Graham et al., 2006; Rapcsak et al., 2000]. Neuropsycholog-ical research suggests that impairments in fear recognition

r Rotshtein et al. r

r 1090 r

may arise only after bilateral lesions to the amygdala[Adolphs et al., 1994], but this is not observed in all bilat-eral amygdala patients [Adolphs et al., 1999; Grahamet al., 2006; Hamann and Adolphs, 1999]. Unilateral amyg-dala lesions are usually associated with intact explicit fearperception [Adolphs et al., 1994], though both unilateralright [Meletti et al., 2003] and left amygdala damage havebeen reported to have some impact on fear perception[Graham et al., 2006]. It is notable that most patients whoshow clear deficits in explicit perception of fear haveextensive lesions not confined to the amygdala but impact-ing also on adjacent structures [Adolphs and Tranel, 2003;Adolphs et al., 2001; Anderson and Phelps, 2000; Ander-son et al., 2000; Brierley et al., 2004; Broks et al., 1998]. Fur-thermore, several patients who show impairment inexplicit perception of fearful expressions due to amygdalapathology were diagnosed with the rare, genetically deter-mined, Urbach–Wiethe syndrome [Adolphs et al., 1994,1995; Siebert et al., 2003]. Neuroimaging studies provideinconclusive evidence regarding the role of the amygdalain explicit recognition of expressions. Some studies suggestthat amygdala activation to fearful stimuli arises primarilyduring explicit recognition tasks [Krolak-Salmon et al.,2004], but many others indicate that amygdala responsive-ness does not depend entirely on explicit expression recog-nition [Bleich-Cohen et al., 2006; Critchley et al., 2000;Gothard et al., 2007; Vuilleumier et al., 2001].

The aims of our study were to determine the time-course of amygdala influences on fearful expression proc-essing, by measuring ERPs to fearful versus neutral faces,and comparing groups of participants with or withoutamygdala damage. We were also interested in determiningany impact of such amygdala damage on explicit fear per-ception from facial expressions. To address thesequestions, we combined neuropsychological and electro-physiological methods. We recorded ERPs to fearful andneutral facial expressions from unoperated patients thatsuffer from medial temporal lobe epilepsy (MTLE), withvarying degrees of amygdala pathology, plus healthy vol-unteers. Our analysis focused on differential (fear minusneutral) effects of emotional expression on evokedresponses, from early stimulus onset up to 600 ms. In anadditional behavioral experiment, we measured explicitrecognition of fearful expressions in the MTLE patientsand healthy participants.

PROCEDURES AND METHODS

Participants included 17 MTLE patients, of whom 7 (4female, 3 left handed, mean age 34.5 years, range 21–42years) had structural pathology that included the amyg-dala (‘‘MTLE-amygdala’’) and 10 others (6 female, 2 lefthanded, mean age 37.7 years, range 23–48 years) withdamage sparing the amygdala but affecting the hippocam-pus and/or other temporal regions (‘‘MTLE-control’’).None of the MTLE patients experienced any seizure for atleast 24 h prior to the study. In addition, 13 healthy con-

trols (7 female, 1 left handed, mean age 31.6 years, range20–58 years) were recruited, most of them from amongfriends and relatives of the MTLE patients. None of thehealthy controls had a clinical history of neurological orpsychiatric illness. All participants gave written informedconsent in accordance with local ethics.

The MTLE patients were recruited from a specialist epi-lepsy center and assigned to groups based on clinical diag-nosis. This diagnosis was made by a clinician blind to ourERP and behavioral hypotheses. Note that all patientswere diagnosed with unilateral brain pathology. Magneticresonance (MR) T2-weighted images were acquired as partof the standard clinical diagnosis, as an indication for scle-rosis [Bartlett et al., 2002]. T2 relaxation time above 92 mswithin the amygdala indicates abnormal tissue [Bartlettet al., 2002]. Furthermore, the distribution of T2 signalwithin the healthy population is skewed and hence highT2 values, even below 92 ms (the clinical threshold), arerare in healthy individuals [Bartlett et al., 2002]. We usedthe structural T2 values as a further regressor in our ERPanalyses, to perform quantitative tests of any relationshipbetween the severity of amygdala pathology and observedERP effects. Six MTLE-amygdala patients were classifiedas showing amygdala sclerosis (five left and one right).One MTLE-amygdala patient was diagnosed as havingdamage to the left amygdala based on assessment of theamygdala structure in her T1-weighted MR images. Thispatient did not suffer from sclerosis.

Importantly, the two patient groups did not differ signif-icantly on clinical measures, including duration of epi-lepsy, seizure severity, treatment, or general intelligence(for details, see Table I), with any nonsignificant trendssuggesting that, if anything, the MTLE-controls wereslightly more impaired than the MTLE-amygdala group.Note that this tendency could only work against the effectswe report below. From all the participants who underwentthe ERP study, 10 healthy, 9 MTLE-control, and 7 MTLE-amygdala participants also completed a behavioral test ofexplicit perception for facial expressions (see below). Cru-cially, during the data collection, the experimenters wereblind to the specific group-assignment of the patients. Theparticipants were also naı̈ve as to the aim of the experi-ment, but were debriefed at the end of the experiment.

Behavioral Study—Explicit Fear Perception

for Seen Faces

Expression perception was evaluated using a categoriza-tion task of prototypical [Ekman and Friesen, 1976] andambiguous facial expressions from the ‘‘morphed-hexa-gon’’ stimulus set [Calder et al., 1996]. The latter stimuliwere included as they are considered to depict more life-like expressions due to their relatively ambivalent natureand hence may be more sensitive for detecting any differ-ences in explicit expression perception [Adolphs and Tra-nel, 2004].

r ERPs for Faces Following Amygdala Damage r

r 1091 r

Stimuli

We used 10 identities, posing six basic prototypicalexpressions (i.e., fear, surprise, happy, angry, disgust,and sad), taken from the Ekman and Friesen set [Ekmanand Friesen, 1976]. Faces were cropped to exclude outlinefeatures and were presented on a gray background. Ofthese, six identities depicting the prototypical expressionswere used to assess the perception of prototypical exem-plars. Morphed images of the remaining four identitieswere used for the morphed-hexagon set. The morphedimages were adapted from Calder et al., [1996]. Therewere five levels of morphing in steps of 20%. Pairing ofexpressions was based on a confusion matrix obtainedfrom healthy volunteers [Calder et al., 1996]. The morph-ing between expressions followed the order: fear–sur-prise–happy–anger–disgust–sad–fear. Each identity wasmorphed within itself. A total of 120 morphs were used(4 identities � 6 expressions � 5 morph levels). Note thatambiguity arises because each morphed face conveys amixture of expressions (see Fig. 1A for examples).

Procedure

Each participant categorized the prototypical exem-plars first: the six identities posing the six prototypicalexpressions. Each stimulus was presented once in ran-dom order with no time limit for a response. The writtennames of the six expressions corresponding to numbers(1–6) were presented below the face image. Assignmentof names to numbers was random across participants.Observers were instructed to press the correspondingnumber (on a keyboard) for the expression that bestdescribed the image, as accurately and quickly as possi-ble. No feedback was provided. Testing with themorphed-hexagon set followed the same procedure andwas run subsequently. Each morph was presented twicein two separate sessions, with a short break in between(total of 240 morphs, 8 for each of the 30 levels).

Analysis

Our analysis for the prototypical exemplars scored cor-rect and incorrect responses to assess fear recognition.We used two approaches to analyze responses to themorphed expressions. The first analysis examinedresponses to morphs that included a degree of fearexpression. A three-way mixed ANOVA was used tocompute effect of group and morphing parameters onthe fear responses. In the second analysis, we computedthe participants’ ‘‘bias’’ in perceiving fearful expressions.The morphed nature of these stimuli provides ambigu-ous expressions, allowing us to determine the tendencyto read fear in the morphed expressions, rather thanscoring the responses as correct/incorrect. This ‘‘fearbias’’ was estimated by computing the relative number offear responses to all the morphed expressions. We

TABLEI.Participantdetails

Group

NSex

HAge

Lesion

Etiology

OnsetEpi

(years)

Frequen

cy(seizu

res/

month)

Med

ication

(mg/day

)VIQ

PIQ

MTLE-control

10Fem

ale¼

5RH

¼7

37.7

(23–48)

LH

(9);

TLE(1)

MTS(9)

5.8(0–14)

8.7(0–7)

2,782

(335–3,400)

90.25

(74–110)

94(68–139)

MTLE-amygdala

7Fem

ale¼

4RH

¼5

34.5

(21–42)

LHA(5);

LA

(1);RA

(1)

MTS(5)

8.5(0.5–17)

3.2(1–35)

1,475

(500–4,350)

100

(83–116)

107

(90–128)

Healthy

13Fem

ale¼

7RH

¼12

31.6

(20–58)

Meansaregiven

withranges

below

ineach

cellwhereap

propriate.

Therewerenosignificantdifferencesbetweenpatients

groupsonan

yclinical

measu

re,exceptlesion

(see

T2

structuralMR

data

inFig.2E

scatter-plot,

along

x-ax

isthere),an

dno

significantdifference

between

patientgroupsan

dhealthy

controls

forag

e,gen

der,or

han

ded

ness.

H,han

ded

ness;

RH,righthan

ded

;H,hippocampus;

HA,hippocampusþ

amygdala;

TLE,temporallobeep

ilep

sywith

exactlesion

unsp

ecified

;MTS,med

ialtemporal

sclerosis

Epi,ep

ilep

sy.Note

that

theetiologyof1MTLE-controlwas

non-specifictemporalep

ilep

tic,

with

noclearstructuralab

norm

ality,2MTLE-amygdalahad

tumourin

the

amygdala.

Med

icationsrefers

tothetotaldosesofdrugsapatientreceive.


r 1092 r

concentrate on this measure rather than on absolute cor-rect responses, as previous work indicates that amygdaladamage usually results in positive judgment biases for thecharacter and expressions of unfamiliar individuals[Adolphs et al., 1998]. Such a positive bias effect has alsobeen observed in monkeys after amygdala damage [Ama-ral, 2003]. A one-way ANOVA was used to compare fearresponses between the groups.

EEG Study

Stimuli and experimental procedure

Photographs of 10 individuals with fearful or neutralexpressions, taken from the Ekman series [Ekman andFriesen, 1976], plus 20 photos of houses were used as stim-uli. Houses were cropped to have an elliptical frame,while faces were cropped to exclude outline features

within a similar frame. Stimuli were presented on a blackbackground (Supporting Information Fig. S1A). A factorialdesign with two levels of expression (neutral or fearful)and two face orientations (upright or inverted), plus anadditional condition of upright houses, was employed.The stimuli were presented in random order, each for 300ms, with an interstimulus interval of 2,200 ms. Partici-pants’ task was to detect any immediate repetition of astimulus (one-back task), and such repetitions arose with a�15% probability regardless of stimulus type.

The EEG experiment was divided into eight sessions. Ineach session, all stimuli were presented once (60 stimuli þ�10 target repetitions). To keep the participant motivatedand alert, they received feedback on their performance atthe end of each epoch (i.e. accuracy and reaction times fordetecting occasional immediate repetitions). The experi-ment started with a practice session, which was structuredas an epoch and included presentation of all stimuli. This

Figure 1.

Explicit perception of emotional expression. (A) Exemplars of

ambivalent expressions used in the morphed-hexagon expression

categorization task. Clockwise, starting from the top-left, the

examples correspond to the midpoint morphs between the fol-

lowing: fear and surprise; happy and anger; anger and disgust; dis-

gust and sad; surprise and happy; sad and fear. Note that the

categorical description of these morphs is unclear and hence influ-

enced by and revealing of any biases in a particular perceiver. (B)

‘‘Fear bias’’ derived for each participant from their explicit judg-

ments of the morphed stimuli, calculated as the ratio of the total

number of fear responses to the total number of morphed expres-

sions shown. All individuals from the healthy (black outline

circles), MTLE-control (green filled circles), and MTLE-amygdala

(red solid triangles) groups are plotted, and no differences were

found between groups. The horizontal lines in the plot depict the

mean score of the healthy participants �2 standard deviations.

(C) The number of ‘‘fearful’’ responses (y-axis) for each morph

type along the expression-hexagon (x-axis). In the expression-hex-

agon, morphs are between two expressions; the x-axis presents

them in following order: surprise–fear, fear–sad, sad–disgust, dis-

gust–angry, angry–happy, and happy–surprise. Note that for all

three groups the number of fear responses peaked for the

morphs: 10% surprise–90% fear and 90% fear–10% sad. MTLE-cnt

(control); MTLE-amyg (amygdala); and CI, confidence interval.

[Color figure can be viewed in the online issue, which is available

at www.interscience.wiley.com.]


r 1093 r

protocol enabled us to familiarize subjects with the set ofstimuli used in the experiment and to minimize any nov-elty effects for the ERP signal. Observers were instructedto maintain fixation throughout the experiment, to avoidhead movements and blinks, and to respond as quicklyand accurately as possible. Stimuli were presented usingCogent1.24 (www.vislab.ucl.ac.uk/Cogent). Prior to theexperiment, a photometer was used to measure the exacttiming of each stimulus onset on the computer monitor, toensure optimal synchronization of the stimulus onsetswith the EEG trigger recording that was generated by theexperimental computer.

EEG recording

Recordings were made with Ag–AgCl electrodes usingthe NeuroScan system and the 10–20 montage system.Twenty-three sites were recorded: Fpz, F7, F3, Fz, F4, F8,FC5, FC6, P7, C3, Cz, C4, P8, CP5, CP6, T5, P3, Pz, P4, T6,Ol, Or, and Oz, with linked-earlobe reference (see Support-ing Information Fig. S2B). Horizontal EOG (HEOG) wasrecorded bipolarly from the outer canthi of both eyes.Electrode impedance was kept <5 kX. Amplifier bandpasswas 0.1–40 Hz. EEG and HEOG were sampled with a digi-tization rate of 200 Hz.

Data analysis

The ERP analysis was performed using SPM5 (WellcomeTrust Centre for Neuroimaging, UCL, London) and Mat-lab7.1. The advantage of using SPM5 is that it provides ameans to correct for familywise errors (FEW) based on therandom field theory [Penny et al., 2003], while using estab-lished statistical methods to test for common and dissoci-ated effects. It thus allows a comprehensive analysismethod not restricted solely to specific electrodes or com-ponents of prior interest. Data preprocessing includedepoching the data from �100 ms prestimulus onset toþ600 ms poststimulus onset. We defined six event types: 2facial expressions � 2 face orientations, houses, and imme-diate-repetition (target) trials. ERPs for each event werecomputed relative to the prestimulus baseline (�100 to 0ms). An absolute threshold for artifact-removal was set to70 lV to exclude events involving eye blinks, lateral orvertical eye movements, and any other artifacts causingdistortions in the EEG. There were no significant differen-ces between the groups in the quality of the EEG signal(see Supporting Information Table S1), and none of thepatients had interictal discharges during the recording.

Importantly, all follow-up statistical analyses were per-formed as interactions of the within-subject and between-subject (group) factors. This approach entails that any non-specific impact of group per se on the ERP signal, whichmight be due to abnormal brain structures, cannot by itselfexplain the more specific results we describe below. Ourgroup comparisons did not assume equal variance.

SPM-ERP statistical analyses were initially implementedon 2D maps generated by a spatial linear interpolation ofeach ERP at each time point [Kiebel and Friston, 2004a,b].We first aimed to replicate the known ERP face effects inall three groups, by contrasting faces versus houses at twotime windows: 100–150 ms, i.e., covering the P1 time win-dow [Liu et al., 2002], and 150–200 ms, i.e., encompassingthe N170 [Bentin et al., 1996; George et al., 1996]. The mainanalysis then focused on expression effects between 100and 600 ms poststimulus onsets using the following succes-sive time windows: 100–150, 150–200, 200–250, 250–300,300–400, 400–500, and 500–600 ms. To allow inferences atthe population level, a second-level random-effects analysiswas performed, where subjects were treated as random var-iables and the independent variable was the averaged dif-ferential effect size (the contrast image) for that window.

At each time window, we first computed an F test of thethree-way interaction of group-by-facial-expression-by-ori-entation across the 2D interpolated images. When thisinteraction was significant, we next performed a separateANOVA for each face orientation condition. Importantly,an effect of amygdala pathology on the ERPs was consid-ered only if the MTLE-amygdala patients significantly(uncorrected P < 0.01 ¼ FWE P < 0.08) differed from boththe control groups (healthy and MTLE-control), thusallowing a replication of the effect, while also controllingfor epileptic condition. Furthermore, to ascertain a causalinfluence of amygdala pathology on emotional responsesat specific time windows, we also computed parametricstatistics where the amygdala pathology was characterizedquantitatively (based on the separately measured struc-tural T2 signal intensity) rather than categorically. The pre-diction was that if the amygdala is directly involved inmodulating ERPs to emotional faces, then more severeamygdala pathology would be associated with reducedemotional effects on the ERP.

In addition to test for any effects of facial expressionsand orientation that were in common for all the threegroups, we also computed a conjunction analysis (with in-termediate null hypothesis [Friston et al., 2005]) for thetwo-way interaction of expression-by-orientation. Thisensured that the reported common effects were evident ineach group separately, with a strength that was largerthan a minimal T value (minT, see ‘‘Results’’). When sig-nificant conjoint interactions were observed, we computedseparate statistical tests for each face orientation. Effects ofexpressions that were independent of epileptic conditionor amygdala pathology were considered as commoneffects only if all three groups displayed this effect sepa-rately (at uncorrected P < 0.05).

To further verify our critical results, we also performeda complementary statistical analysis independent of theSPM approach (i.e. no longer on the 2D interpolatedmaps). High resolution time-bin analysis (5 ms) was per-formed to give a more fine-grained temporal characteriza-tion, at each electrode separately. Two-sample (notassuming equal variance) and one-sample t-tests were


r 1094 r

used to compute the reliability of differential fear versusneutral effects between and across groups, respectively,for successive time-bins. A threshold of P < 0.05 and atemporal extent of 20 ms were used for inferences andapplied to the resulting statistical ‘‘maps’’ of ERP effectsacross the narrow time-bins.

RESULTS

Behavioral Measures

Explicit fear perception

There were no significant differences between groupsfor correct explicit recognition of the prototypical fearfulexpressions (all P > 0.5). Proportion of accurate responses(mean � SD) was as follows: healthy ¼ 0.61 � 0.23;MTLE-control ¼ 0.61 � 0.28; MTLE-amygdala ¼ 0.52 �0.37. Mixed ANOVA was used to analyze participants’responses to the morphed stimuli that included somedegree of fear expressions (see Fig. 1). The design includedgroup as a between factor (e.g. healthy, MTLE-control,MTLE-amygdala) and two within factors: the expressionthat was morphed with fear (sad, or surprise) and percent-age of fear in the morphed expression (i.e. 10, 30, 50, 70,or 90%). Neither amygdala damage nor epileptic conditionaffected the responses to the morphed fear expressions (allP > 0.1). More detailed behavioral results are reported inthe Supporting Information Results.

Likewise, the fear-bias measure from the morphed-hexa-gon stimulus set indicated no significant differencesbetween the three groups, i.e. healthy, MTLE-amygdala,MTLE-control (all P > 0.15). Out of a total of 240 ambigu-ous expressions, healthy participants categorized 16.7% �5 (SD) as fearful; for MTLE-control patients, this was12.4% � 6; and for MTLE-amygdala patients, 17.7% � 5.8;see Figure 1B for individual performance.

One back task

In the one-back task performed during ERP recording(see ‘‘Procedures and Methods’’), patients and healthy par-ticipants’ responses did not differ (all Ps > 0.1; see Sup-porting Information for details and SupportingInformation Fig. S1). Reaction times (RTs) were notaffected by any of the experimental conditions, neitherstimuli type nor group (all Ps > 0.1). There were also nodifferences in accuracy between faces and houses. Overall,all groups were more accurate in detecting repetitions ofupright than inverted faces (F1,26 ¼ 9.97, P < 0.01) andrepetition of fearful than neutral faces (F1,26 ¼ 5.1, P <0.05), but these two factors did not interact (P > 0.1). Thelack of RT differences between conditions and also of anymain effects or interactions involving the group factor,even for accuracy, suggests that the ERP differencesreported below could not have been confounded by taskor attentional demands. We note also that our ERP analy-ses did not include the repeated trials, and so error-rate

for those on target repetition trials could not affect theresults for the nontarget trials’ analyses.

EEG Data

Category-selective responses, faces versus houses

We first assessed general visual processing in the threeparticipant groups. House stimuli were included as a con-trol visual category to probe the processing of neutral non-social stimuli via ERPs, as compared with faces.Differential evoked responses to faces versus houses didnot differ between the three groups (Supporting Informa-tion Results and Supporting Information Fig. S2). All threegroups showed the expected N170 effect for faces com-pared with houses that peaked at electrode P8 (all Ps <0.05), and also the delayed response to inverted faces com-pared with upright faces at this time window (all Ps <0.05). Furthermore, healthy subjects and MTLE-amygdalapatients also showed an earlier face versus house effectwithin the usual P1 time window (P < 0.05; SupportingInformation Results and Supporting Information Fig. S2).There was no significant interaction of the P1 face effectwith group at this time window (P > 0.1). These ERPresults indicate that in our experiment neither amygdalapathology nor medical condition (i.e. presence of temporallobe epilepsy) affected face-specific processing, as meas-ured by face-versus-house or upright-versus-inverted-faceERP differences. Instead, significant effects of amygdalapathology upon ERPs, as reported below, were specific tothe comparison of facial expressions.

Emotion effects: Fearful versus neutral faces in

the early P1 time window

Most critically, at the time window of 100–150 ms, corre-sponding to the P1 component, we observed a significantthree-way interaction of group-by-expression-by-face ori-entation, peaking at electrode Oz (Z ¼ 2.32, P ¼ 0.01, seeSupporting Information Fig. S3A). Accordingly, we nextanalyzed ERPs for each face orientation separately. Therewere no significant expression and group effects forinverted faces (all Ps > 0.1). This indicates that the P1expression effects reported below cannot be attributed toany low-level feature differences between expressions thatwould be shared between upright and inverted faces (seealso Eimer and Holmes [2002]). In contrast, for uprightfaces a significant interaction of group-by-expression wasobserved, involving lateral posterior electrodes on bothsides, with a peak in vicinity to electrode Or (Z ¼ 2.11,P ¼ 0.018). This interaction reflected differences in the feareffect for MTLE-amygdala patients versus healthy subjects(peaking in vicinity to Or, Z ¼ 2.6, P ¼ 0.003; see Fig. 2),and importantly also for MTLE-amygdala versus MTLE-control patients (peaking in vicinity to P7, Z ¼ 2.37, P ¼0.009; see Fig. 2).

For the interaction described above, at this early timewindow, both the healthy group and the MTLE-control


r 1095 r

patients showed a significantly greater positivity for fear-ful than neutral faces. This emotional positivity enhance-ment was maximal over posterior electrodes, with adistributed scalp effect (Fig. 2A,B). For the healthy group,such an expression-effect on P1 was observed bilaterally atseveral posterior electrodes, maximal at Ol and Or (Z ¼

1.91, P ¼ 0.028, see Fig. 2D); while for the MTLE-controlpatients, this effect predominated over the left-side andpeaked at the nearby electrode T7 (Z ¼ 2.34, P ¼ 0.01).Direct comparison showed that these latter two groupsdid not differ significantly in the 100–150 ms timewindow.

Figure 2.


r 1096 r

By contrast, for the MTLE-amygdala patients, no elec-trode showed any reliable increased positivity to fearfulversus neutral expressions, over the whole scalp, duringthe same time window (all Ps > 0.1). Thus, the increasedpositivity triggered by fearful expression in the 100–150ms time window, which was found in both the healthyand MTLE-controls participants, was eliminated by pathol-ogy in the MTLE-amygdala group. A complementaryanalysis based on successive 5 ms time windows for eachelectrode separately (see Fig. 3A) revealed that a signifi-cant differential (fear-related) response between groupsarose at occipital electrodes (e.g. Oz), starting at 120 mspoststimulus onset.

If the amygdala is indeed critical for modulating theearly (100–150 ms, P1) visual processing of fearful faces,then in addition to the group effect described above, theactual severity of amygdala pathology in individualsshould predict the degree of ERP attenuation of the feareffect in this time window. Severity of amygdala pathol-ogy was defined here based on the separately measuredstructural T2 relaxation time (see ‘‘Procedures and Meth-ods’’). We tested for any relation of the MR T2 signal withthe magnitude of P1 modulation by emotion expression,using the same parametric approach used in a previousfMRI study of temporal sclerosis patients [Vuilleumieret al., 2004], but now applied to ERP data instead. As pre-dicted, we found significant negative correlations of boththe left and right amygdala T2 signal with the expressioneffect on ERPs (upright fearful minus neutral), whichpeaked in vicinity to electrode P7 (for left amygdala struc-tural T2 signal: Z ¼ 2.2, P ¼ 0 014; for right amygdalastructural T2 signal, Z ¼ 1.54, P ¼ 0.061). There were nosignificant differences between right and left amygdala

effects (all Ps > 0.2). These correlations demonstrate thatthe worse the amygdala pathology, the greater the attenua-tion of the early P1 modulation by fearful versus neutralexpressions in ERPs (Fig. 2E).

We note that these correlations were observed even forthe right amygdala that showed only subclinical pathologyaccording to MR T2 signal. This P1 correlation with theright amygdala seemed to be driven to a large extent bythe single patient that had, radiologically, significant rightamygdala damage, as removing this patient from the anal-ysis reduced the extent of the correlation (r ¼ �0.332, Z ¼1.2, P ¼ 0.13).

Emotion effects: Fearful versus neutral faces

in the midlatency N1 and N2 time windows

In the two time windows that followed, 150–200 and200–250 ms, there was no significant three-way interactionof group-by-expression-by-orientation (P > 0.1). Instead,the conjunction analysis revealed that all three groupsshowed in common an interaction between expression andorientation (Supporting Information Fig. S3B). These inter-action effects across group overlapped in vicinity to elec-trodes FC5, FC6, and P4 (Z > 2.21, minT6 > 1.22). Testingeach face orientation separately revealed that none of thegroups showed a significant expression effects withinverted faces (all Ps > 0.1). However, all three groupsshowed an expression effect for upright faces that was pri-marily observed over the central frontal electrodes. Themaximal effects common for the three groups arose in vi-cinity to electrode FC5 at 150–200 ms (Z ¼ 2.64, P ¼ 0.004)and at 200–250 ms (Z ¼ 2.38, P ¼ 0.009; see Fig. 4), with nosignificant group-by-expression interactions (all P > 0.1).

Amygdala damage effects at 100–150 ms. (A) 2D topographic maps

of scalp distribution (occipital regions appear at the bottom of each

scalp map), depicting ERP responses at 115 ms poststimulus onset.

The leftmost column shows responses to upright neutral faces and

the next column the responses to upright fearful faces, for each of

the three groups. Warmer colors represent positive ERPs. Note

that healthy (upper row) and MTLE-control patients (middle row)

show an increased positivity for fearful faces compared with neutral

faces, most notably in posterior electrodes, while MTLE-amygdala

patients do not. This becomes even more evident in (B), which

shows 2D interpolated maps separately for each group, depicting

the fearful-minus-neutral subtraction for upright faces, averaged

across 100–150 ms poststimulus onset. Warm colors represent

more positive ERP responses for fearful than neutral upright faces.

(C) SPMs thresholded at P < 0.05, representing a significant inter-

action of expression (fearful minus neutral upright faces) by group.

(D) Grand-averaged ERP waveform for upright fearful (red) or neu-

tral (dotted black) faces, at electrode Ol, plotted separately for

each group. The P1 label highlights the early ERP expression effect.

Note that the two control groups show a more positive P1 compo-

nent for fearful than neutral faces, while in the MTLE-amygdala

group (shown in the rightmost graph) this positivity is diminished

or even tends to be smaller for fearful than neutral faces. (E) Cor-

relation between the severity of amygdala structural-abnormality

(as separately measured by T2 imaging) and the size of the expres-

sion ERP effects (fear minus neutral for upright faces) during the

100–150 ms time window. The left panel presents an SPM depicting

pixels in the interpolated scalp image that show a conjoint signifi-

cant correlation (P < 0.05) of the left and right structural T2 signal

from amygdala with the P1 expression effect. The plots on the right

shows the T2 signal of right or left amygdala plotted against the

expression effect on P1, extracted from the topography peak (in vi-

cinity to P7 for the left amygdala and P3 for the right amygdala).

Red circles depict the MTLE-amygdala patients and black circles the

MTLE-control patients. Note that more severe damage to the

amygdala is associated with smaller expression effect (fearful � neu-

tral) in ERPs for the 100–150 ms time window. The correlation of

the right amygdala seems to be primarily driven by the single patient

who had a radiologically evident lesion to this structure. MTLE-

amyg ¼ patient group with amygdala damage; MTLE-cnt ¼ control

group of temporal-lobe-epilepsy patients, with structurally intact

amygdala.

Figure 2.


r 1097 r

When considering each group separately for the 150–200ms interval, the maximal effect was at electrode Cz in thehealthy (Z ¼ 2.77, P ¼ 0.003), at T7 in MTLE-controlpatients (Z ¼ 2.26, P ¼ 0.012), and at F8 in MTLE-amyg-dala patients (Z ¼ 2.16, P ¼ 0.015), although similar effectswere seen at the neighboring electrodes and overlappedbetween groups. Likewise, for the 200–250 ms interval,fearful minus neutral effects peaked at electrode F8 forhealthy (Z ¼ 3.66, P < 0.001), FC6 for MTLE-control in vi-cinity (Z ¼ 2.2, P ¼ 0.012), and F7 for MTLE-amygdala invicinity (Z ¼ 2.05, P ¼ 0.02).

Unlike the earlier time window of 100–150 ms, the levelof amygdala structural abnormality did not correlate withthe fear effects for these two intervals, neither at 150–200nor at 200–250 ms (all P > 0.1). The preserved fear effectsduring these time windows in both patient groups werefurther confirmed by testing successive 5 ms time win-dows for each electrode separately (Fig. 3B).

Emotion effects: Fearful versus neutral faces forlater ERP components

At 250–300 ms, a three-way interaction of group-by-expression-by-orientation was observed, now in vicinity toelectrode T7 (Z ¼ 1.94, P ¼ 0.025; see Supporting Informa-tion Fig. S3A). However, this interaction reflected epilepticcondition rather than amygdala pathology in particular.For this time window, the two MTLE groups differedfrom the healthy group in their brain responses to facialexpressions and orientations. As this was not clearly

related to amygdala function in particular and so fallsbeyond the focus of the current research, we did not pur-sue further our analysis for this time window. We notealso that the finer time-bin analysis (Fig. 3A) suggestedonly weak and marginal amygdala effects around 280 mspoststimulus onset, involving Or and P4 electrodes only.As these effects were marginal and did not produce a criti-cal three-way-interaction nor a group-by-expression inter-action when tested for upright faces only (all Ps > 0.1), wedid not consider them further.

The next significant ERP effects were observed at 300–400 ms (Supporting Information Fig. S3B). Here all threegroups showed a significant expression-by-orientationinteraction, with a conjoint maxima peak in vicinity to F7(Z ¼ 2.68, P ¼ 0.004, minT6 ¼ 1.6). Further analysisrevealed a significant conjoint effect of expression forupright faces, peaking in vicinity to F8 (Z ¼ 1.92, P ¼0.027) and F7 (Z ¼ 1.67, P ¼ 0.048), but no conjoint effect ofexpression with inverted faces (P > 0.05). All groupsshowed increased positivity for fearful compared to neutralexpressions in upright faces, maximal in vicinity to F7 inhealthy (Z ¼ 2.51, P ¼ 0.006), to P4 in MTLE-control (Z ¼1.36, P ¼ 0.087), and to F8 in MTLE-amygdala (Z ¼ 1.42, P¼ 0.078) groups. We note that this effect was most reliablein the healthy participants, indicating that for this time win-dow, epileptic condition rather than specific amygdala pa-thology might exert some (marginal) influence.

Finally, three-way interactions (group-by-expression-by-orientation) were also observed for the 500–600 ms timewindow, in vicinity to P8 (Z ¼ 1.61, P ¼ 0.05, see

Figure 3.

Time-course of expression effects. Statistical parametric results of

point-by-point t-tests, with the y-axis corresponding to particular

electrodes, and the x-axis to particular successive 5 msec time-

bins post-stimulus. The statistical maps are threshold at P < 0.05

and depict the detailed time-course for expression effects (upright

fearful minus neutral) that last longer than 20 msec. (A) Time-

bins (in hot colours) during which MTLE-amygdala patients signifi-

cantly differed from MTLE-control patients and healthy volun-

teers, in response to the different upright expressions. Amygdala

related abnormalities in ERP responses can readily be seen in two

time-windows: An early effect, arising at around �100 msec and

lasting till �150 msec is mostly expressed in posterior electrodes;

and a later effect (�500–600 msec) is expressed more widely in

central-posterior electrodes. Note that the later effect was mostly

pronounced at around 500 and 575 msec. (B) Time-bins during

which all three groups showed similar expression effects (for

upright fearful minus neutral faces), that were unaffected by amyg-

dala damage. Around �180 msec post stimulus onset, at frontal

central electrodes, a main effect of expression (upright fear >neutral) was observed, with a similar pattern across all three

groups. Y-axis, individual electrodes (top-to-bottom): F, frontal:

FPZ, F8, F4, FZ, F3, F7, FC6, FC5; T, Temporal: T8, T7; C, central:

CP6, C4, CZ, C3, CP5; P, Parietal: P8, P4, PZ, P3, P7; O, occipital:

OR, OZ, OL. [Color figure can be viewed in the online issue,

which is available at www. interscience.wiley.com.]


r 1098 r

Supporting Information Fig. S3A). Separate analysisrevealed, for upright faces only, a significant group-by-expression interaction in vicinity to P8 (Z ¼ 2.54, P ¼0.006). The MTLE-amygdala differed from the healthy atcentral posterior electrodes with a maxima at P8 (Z ¼ 2.71,P ¼ 0.003) and from the MTLE-control with a maxima alsoat P8 (Z ¼ 3.11, P ¼ 0.001 see Fig. 5). The two groups withintact-amygdala did not differ significantly (P > 0.1). Dur-ing this late time interval, healthy and MTLE-controlpatients, but not MTLE-amygdala (P > 0.1), showed anincreased positivity for fearful versus neutral upright

expressions at several posterior central electrodes (healthypeaking in vicinity to Ol, Z ¼ 2.8, P ¼ 0.003; MTLE-con-trols peaking in vicinity to P4, Z ¼ 2.48, P ¼ 0.007). Thisoutcome was further confirmed by our finer-grained tem-poral analysis using successive 5 ms time windows foreach electrode separately (Fig. 3A), which confirmed aclear amygdala-dependent difference around 500–600 ms.

Moreover, the extent of separately measured, structuralamygdala pathology (measured as the T2 MR signal) againcorrelated negatively with the size of the fear effect for the500–600 ms window, with more severe amygdala damage

Figure 4.

Expression effects at 150–250 ms that were not affected by

amygdala damage. (A) 2D topographic maps of scalp distribution

depicting ERP responses at 185 ms poststimulus onset. The left-

most column represents responses to upright neutral faces and

the next column shows responses to upright fearful faces.

Warmer colors represent positive ERPs. Note that all three

groups, healthy (upper row), MTLE-control patients (middle

row), and MTLE-amygdala patients (bottom row), showed an

increased positivity for fearful faces compared with neutral faces

at this time. (B) 2D interpolated maps shown separately for

each group, for the subtraction of upright fearful minus neutral

faces, averaged across the time window 150–200 ms poststimu-

lus onset. Warmer colors represent more positive ERP

responses for fearful than neutral expressions. (C) SPMs thresh-

old at P < 0.05, presenting the conjunction of expression effects

(upright fear > neutral) found in common across all three

groups, for the two successive time windows of 150–200 and

200–250 ms poststimulus onset. There were no interactions of

expression with group in these time windows and thus no

impact of amygdala damage or of temporal-lobe epilepsy. (D)

Averaged ERP waveforms for upright fearful (red) or neutral

(dotted black) faces at electrode Fz, plotted separately for each

group. The ellipses highlight the 150–250 ERP expression effects,

as shown by each group in common.


r 1099 r

leading to more pronounced attenuation of fear-relatedpositivity. The conjoint effect of left and right amygdaladamage peaked in vicinity to P8 (Z ¼ 2.47, P ¼ 0.007,minT16 ¼ 1.46). The right amygdala parametric modula-tion of the upright fear > neutral responses during thistime window peaked in vicinity to P4 (Z ¼ 2.62, P ¼0.004), while the parametric modulation due to left amyg-dala pathology peaked in vicinity to P8 (Z ¼ 1.75, P ¼0.04). In this late time window, frontal electrodes alsoshowed some effect of right (but not left) amygdala pa-thology, with peak differences in vicinity to electrode F7(Z ¼ 2.45, P ¼ 0.007). This differential effect of right amyg-

dala pathology is intriguing given the subclinical pathol-ogy of MR signal for the right amygdala in all MTLEparticipants but one. However, similar to the effects forthe earlier P1 effect (Fig. 2E), the present correlation withthe right amygdala seemed to be substantially driven bythis single patient with a radiologically evident rightamygdala lesion (Fig. 5C), as removing this patient fromthe analysis reduced the extent of the correlation (r ¼�0.34, Z ¼ 1.18, P ¼ 0.12) for the late 500–600 ms timewindow.

In this late time window, an interaction between expres-sion and orientation was also found to produce a bilateral

Figure 5.

Amygdala abnormality effects at 500–600 ms. (A) SPMs thresh-

olded at P < 0.05, for the significant interaction of expression

(upright fearful minus neutral) by group in the 500–600 ms time

window. (B) Averaged ERP waveforms for upright fearful (red)

or neutral (dotted black) faces at electrode P4, plotted sepa-

rately for each group. The ellipses highlight the late ERP expres-

sion effect that differs between groups. Note that similar to the

results in Figure 3A, larger differences arose at around 500 ms

and again at around 575 ms at peristimulus time. (C) Correla-

tion of the severity of amygdala structural-abnormality and the

size of the ERP expression effects (upright fear minus neutral) in

the 500–600 ms time window. The left panel shows an SPM

depicting pixels in the interpolated scalp image that show a con-

joint significant correlation (P < 0.05) of the left and right struc-

tural T2 signal from amygdala with the late expression effect.

The plots on the right shows the T2 signal of right or left amyg-

dala plotted against the expression effect extracted from the

peak (in vicinity to P8). Red circles depict the MTLE-amygdala

patients and black circles the MTLE-control patients. Note that

the correlation of the right amygdala seemed to be driven

mainly by the single patient who had a radiologically evident

lesion to this structure. MTLE-amyg ¼ patient group with amyg-

dala damage; MTLE-cnt ¼ control group of temporal-lobe-epi-

lepsy patients, with structurally intact amygdala. [Color figure

can be viewed in the online issue, which is available at

www.interscience.wiley.com.]


r 1100 r

conjoint effects across all groups, peaking in vicinity to F7(Z ¼ 2.16, P ¼ 0.015; see Supporting Information Fig. S3B).Intriguingly, this conjoined frontal effect originated froman expression effect observed for inverted faces in particu-lar. Here, all three groups showed increased negativity forinverted fearful compared to inverted neutral faces (Z ¼1.77, P < 0.038), but not for upright faces (P > 0.1). Thisexpression effect may reflect different visual processingdemands for inverted faces, but for present purposes themain point is that this was not affected by amygdalapathology.

DISCUSSION

In this study, we measured EEG responses to threatstimuli in three groups (healthy controls, temporal lobeepilepsy patients with intact amygdale and with damagedamygdala). We show that the human amygdala has a sig-nificant role in modulating brain responses to threat-related facial expressions at specific, early and late, timepoints. Critically, we show that the impact of amygdaladamage expresses itself at two specific time windows post-stimulus onset: 100–150 and 500–600 ms. In addition, weobserve differential responses to fearful versus neutralexpressions in upright faces during an intervening timewindow (150–250 ms) that remain intact despite amygdaladamage. We propose that these latter findings may relateto the fact that explicit perception of fearful expressionswas preserved in our MTLE-amygdala participants, asshown by their behavioral responses. By contrast, bothearly and late ERP responses to fearful faces were selec-tively disrupted by amygdala damage, but preserved inMTLE patients with intact amygdala, as well as for healthycontrols. Taken together, the results indicate multiple proc-essing stages or routes for fearful face stimuli, which varyin their timing and their dependence on normal amygdalafunction. Our data reveal for the first time that amygdalapathology disrupts an early and a late cortical processingstage for fearful faces while sparing an intermediate stageof emotion-related processing.

The earliest ERP effect specific to upright fearful versusneutral faces here was observed in healthy and MTLE-con-trol groups within �100–150 ms, emerging as a ‘‘fear’’enhanced positivity at �120 ms after stimulus onset, overposterior electrodes. By contrast, within this same timewindow, an abnormal ERP response in the MTLE-amyg-dala group was observed, corresponding to a loss of thisearly fear-induced positivity (Figs. 2 and 3A). Moreover,the attenuation of this effect was larger in patients withmore severe amygdala structural pathology, suggesting adirect causal role for the amygdala. This early ERP expres-sion effect corresponds with the well-established latencyand spatial distribution of the P1 component, which isthought to arise from extrastriate visual cortex [Di Russoet al., 2002, 2003]. Therefore, these results provide new evi-dence that processing of fearful faces in visual cortex is

susceptible to a rapid modulation by the amygdala. Thissupports previous suggestions that the amygdala providesrapid and ‘‘reflexive’’ feedback to cortical processing, topromote an effective response to potential threat [Amaral,2002; Dolan, 2002; LeDoux, 1996; Phelps and LeDoux,2005; Vuilleumier, 2005].

A second abnormal ERP fearful response in the MTLE-amygdala group was expressed in a much later time win-dow, 500–600 ms poststimulus onset (i.e., correspondingwith the late-P3, Figs. 3A and 5). Emotion-related ERPeffects in this time-range have been associated with possi-ble arousal responses [Kissler et al., 2006], while othershave linked them to enhanced episodic memory-relatedprocessing for emotional stimuli [Maratos et al., 2000].This latter interpretation might appear potentially at oddswith the behavioral results reported here, e.g., no effect ofamygdala pathology on the one-back task. However, wenote that the late-P3 ERP emotional memory effect haspreviously been discussed in the context of long-term epi-sodic memory processing [Maratos et al., 2000] rather thanin the context of short-term working memory as requiredby the one-back task used here. In support of the proposalthat the amygdala may be involved with episodic emo-tional memory processing, it has been shown that the se-verity of structural amygdala abnormality in MTLEpatients predicts mnemonic impairments related to emo-tional stimuli [Richardson et al., 2004], i.e., a loss of theusual long-term memory advantage for emotional relativeto neutral stimuli. A wide literature suggests a crucialinvolvement of amygdala in emotional memory processes[Dolan, 2002; LaBar et al., 1995]. Our new ERP results inthe 500–600 ms (‘‘late P3’’) time window may relate to this.

Taken together, our early (P1) and late (P3) amygdala-damage effects on expression-related ERPs indicates thatthe amygdala directly affects both perceptual–attentionaland mnemonic processing. P1 modulations have oftenbeen reported to relate to attention-related enhancementsof processing in visual cortex [Di Russo et al., 2003;Holmes et al., 2003], and such visual effects might accountnot only for better detection of emotional stimuli in atten-tion tasks [e.g., Phelps and LeDoux, 2005; Vuilleumier,2005] but also potentially contribute to strengthening sub-sequent memory traces for emotional material [Adolphset al., 2005; Buchanan et al., 2006; Talmi et al., 2007]. Ourfindings that amygdala damage also affected a later dis-tinct ERP component (late P3), associated with memory,provides a novel line of evidence for amygdala influenceson learning and memory formation, on top of any earlierattentional effect.

In contrast to the abnormal early and late ERPresponses, the MTLE-amygdala patients demonstrated nor-mal explicit perception of fearful expressions. This wasobserved for prototypical Ekman expressions [Ekman andFriesen, 1976], and also when using a more sensitive mea-sure with the morphed-hexagon expressions [Calder et al.,1996]. These findings suggest that normal functioning ofthe amygdala is not a prerequisite for explicit fear


r 1101 r

recognition, consistent with some other reports [Adolphset al., 1995; Graham et al., 2006; Rapcsak et al., 2000]. Fur-thermore, amygdala pathology did not affect performancein the one-back task performed during EEG recording thatis presumably based on short-term or working memory.All three groups showed more accurate responses to fear-ful compared with neutral expressions in the one-backtask, and this was unaffected by amygdala damage.

Likewise, we also observed preservation of fearfulexpression effects on ERPs, specifically during the 150–250ms (and to a lesser degree the 300–400 ms) time windowsthat intervened between the early and late abnormalities.A fear effect during this time window was observed forthe MTLE-amygdala participants, just as for the healthygroup and the MTLE-controls (Figs. 3B and 4). Anenhanced positivity in response to fearful as compared toneutral faces has been found between 130 and 200 ms afterstimulus onset in several previous ERP studies (see Eimerand Holmes [2007], for a review). This effect is assumed tobe linked to the explicit detection of fearful faces and typi-cally shows a fronto-central scalp distribution, similar towhat was observed in the present experiment during the150–250 ms time interval for all three groups.

In keeping with the explicit behavioral judgments andperformance of the one-back task, these aspects of the ERPdata indicate that not all fear-related processes were dis-rupted in MTLE-amygdala patients. We surmise that thepreserved ERP and behavioral effects might be linked,pointing to a processing stage at around 150–250 ms post-stimulus onset mediating both explicit fear perception andemotional effects on working memory. This hypothesisaccords with previous ERP data in healthy participantsthat tentatively linked fearful ERP effects at this time win-dow to explicit processing of expression [Mikhailova andDavydov, 1999], and showed that these ERP responses toexpressions depend on attention to the stimulus [Eimerand Holmes, 2007]. In any case, the preserved fear effectfor this time-period in MTLE-amygdala subjects certainlydemonstrates that not all processing of fearful facesdepends on contributions from the amygdala, suggestinginstead that there are multiple pathways for processingemotional information [Amaral, 2002; LeDoux, 1996; Vuil-leumier, 2005], each involving differential neural structuresand importantly operating with a specific time course.

Although our ERP data revealed that amygdala pathol-ogy affected fear processing in two distinct time windows,no deficits were evident in the particular behavioral meas-ures used in this study (explicit recognition of expressions,or performances in a one-back memory task). This raisesthe question of what is the possible functional significanceof these ERP effects. We speculate that the early ERPexpression effect (around the P1) might constitute amarker for a signal that acts to direct processing resourcestoward potential threat in the environment, in accord withprevious imaging data on P1 responses to emotional stim-uli [e.g., Pourtois et al., 2004, 2005b]. This neural markercould possibly also prepare the autonomic response sys-

tem to react to potential threat in the environment [Critch-ley et al., 2005; Williams et al., 2004]. We further suggestthat this signal might facilitate the explicit perception ofexpression but is not necessary for it. Therefore, we pre-dict that amygdala damaged patients should showimpaired autonomic responses to threat stimuli andreduced attention-capture by threat-related stimuli (seeAnderson and Phelps [2001]). The late ERP effect (500–600ms) may relate to the facilitation of emotional influenceson encoding into long-term memory. Thus, the sameamygdala patients should show a loss of the usualenhancement in long-term memory for emotional items(see Adolphs et al. [2000] and Richardson et al. [2004]).Further research is needed to establish the exact behavioralsignificance of the different effects of amygdala damageon ERPs to various classes of emotional stimuli, in differ-ent perceptual and memory task conditions. But for pres-ent purposes, the present ERP results clearly establishesan impact of amygdala damage on the brain response tofearful versus neutral faces at distinct time windows,while showing other brain responses remain intact atintervening time windows.

Finally, in accordance with previous literature [Rotshteinet al., 2001], our data suggest that during the first 600 mspoststimulus onset, the amygdala was involved only inprocessing fearful expressions for upright but not invertedfaces. Effects of fearful expression with inverted faceswere observed at 500–600 ms, but crucially those were notaffected by amygdala pathology. This particular late effectmay correspond to a delay in expression recognition thathas been reported as typical for inverted faces [McKelvie,1995]. Further research is needed to gain a better under-standing of the fear effect in inverted faces. In addition, itis interesting to note that medial temporal lobe pathology,irrespective of amygdala involvement or sparing, pro-duced some distinct effects on expression processing at250–300 ms. This dissociation between general MTLEeffects, versus amygdala-specific effects, emphasizes thenecessity of including a control group that shares etiologyand medical history with the neurological group of maininterest (MTLE-control vs. MTLE-amygdala, respectively),as done here, to better isolate the specific relationshipbetween brain structures and function.

In conclusion, our ERP data show for the first time thatthe amygdala makes distinctive contributions to process-ing of fearful faces in at least two distinct time windows,both early (100–150 ms) and late (500–600 ms). We demon-strate a reduction of these effects due to amygdala dam-age, in direct parametric proportion with the structuralseverity of the pathology. Conversely, we show that at anintervening time window (150–250 ms), the effects of fear-ful expression on brain responses were preserved despiteamygdala pathology. We conclude that threat-related stim-uli are processed within multiple neural stages and path-ways, some dependent on the amygdala and others not,each with a particular time-course. More generally, ourstudy illustrates how combining the lesion approach with


r 1102 r

EEG recording can uncover the specific time-points atwhich the damaged area normally contributes to influenceinformation processing in distant cortical regions that sur-vive the lesion.

ACKNOWLEDGMENTS

We thank S. Kennett, A. Quayle, and U. Noppeney forhelp, plus the patients and their relatives for participating.

REFERENCES

Adolphs R, Tranel D (2003): Amygdala damage impairs emotionrecognition from scenes only when they contain facial expres-sions. Neuropsychologia 41:1281–1289.

Adolphs R, Tranel D (2004): Impaired judgments of sadness butnot happiness following bilateral amygdala damage. J CognNeurosci 16:453–462.

Adolphs R, Tranel D, Damasio H, Damasio A (1994): Impairedrecognition of emotion in facial expressions following bilateraldamage to the human amygdala. Nature 372:669–672.

Adolphs R, Tranel D, Damasio H, Damasio AR (1995): Fear andthe human amygdala. J Neurosci 15:5879–5891.

Adolphs R, Tranel D, Damasio AR (1998): The human amygdalain social judgment. Nature 393:470–474.

Adolphs R, Tranel D, Hamann S, Young AW, Calder AJ, PhelpsEA, Anderson A, Lee GP, Damasio AR (1999): Recognition offacial emotion in nine individuals with bilateral amygdaladamage. Neuropsychologia 37:1111–1117.

Adolphs R, Tranel D, Denburg N (2000): Impaired emotional de-clarative memory following unilateral amygdala damage.Learn Mem 7:180–186.

Adolphs R, Tranel D, Damasio H (2001): Emotion recognitionfrom faces and prosody following temporal lobectomy. Neuro-psychology 15:396–404.

Adolphs R, Tranel D, Buchanan TW (2005): Amygdala damageimpairs emotional memory for gist but not details of complexstimuli. Nat Neurosci 8:512–518.

Amaral DG (2002): The primate amygdala and the neurobiologyof social behavior: Implications for understanding social anxi-ety. Biol Psychiatry 51:11–17.

Amaral DG (2003): The amygdala, social behavior, and dangerdetection. Ann N Y Acad Sci 1000:337–347.

Anderson AK, Phelps EA (2000): Expression without recognition:Contributions of the human amygdala to emotional communi-cation. Psychol Sci 11:106–111.

Anderson AK, Phelps EA (2001): Lesions of the human amygdalaimpair enhanced perception of emotionally salient events. Na-ture 411:305–309.

Anderson AK, Spencer DD, Fulbright RK, Phelps EA (2000): Con-tribution of the anteromedial temporal lobes to the evaluationof facial emotion. Neuropsychology 14:526–536.

Armony JL, Quirk GJ, LeDoux JE (1998): Differential effects ofamygdala lesions on early and late plastic components of audi-tory cortex spike trains during fear conditioning. J Neurosci18:2592–2601.

Aru J, Bachmann T (2009): Occipital EEG correlates of consciousawareness when subjective target shine-through and effectivevisual masking are compared: Bifocal early increase in gammapower and speed-up of P1. Brain Res 1271:60–73.

Ashley V, Vuilleumier P, Swick D (2004): Time course and speci-ficity of event-related potentials to emotional expressions. Neu-roreport 15:211–216.

Bartlett PA, Richardson MP, Duncan JS (2002): Measurement ofamygdala T2 relaxation time in temporal lobe epilepsy. J Neu-rol Neurosurg Psychiatry 73:753–755.

Bentin S, Allison T, Puce A, Perez E, McCarthy G (1996): Electro-physiological studies of face perception in humans. J CognNeurosci 8:551–565.

Bleich-Cohen M, Mintz M, Pianka P, Andelman F, Rotshtein P,Hendler T (2006): Differential stimuli and task effects in theamygdala and sensory areas. Neuroreport 17:1391–1395.

Brierley B, Medford N, Shaw P, David AS (2004): Emotional mem-ory and perception in temporal lobectomy patients with amyg-dala damage. J Neurol Neurosurg Psychiatry 75:593–599.

Brisson B, Jolicoeur P (2008): Express attentional re-engagementbut delayed entry into consciousness following invalid spatialcues in visual search. PLoS ONE 3:e3967.

Broks P, Young AW, Maratos EJ, Coffey PJ, Calder AJ, Isaac CL,Mayes AR, Hodges JR, Montaldi D, Cezayirli E, Roberts N,Hadley D (1998): Face processing impairments after encephali-tis: Amygdala damage and recognition of fear. Neuropsycholo-gia 36:59–70.

Buchanan TW, Tranel D, Adolphs R (2006): Memories for emo-tional autobiographical events following unilateral damage tomedial temporal lobe. Brain 129:115–127.

Calder AJ, Young AW, Perrett D, Etcoff NL, Rowland D (1996):Categorical perception of morphed facial expressions. VisCogn 3:81–117.

Carlsson K, Petersson KM, Lundqvist D, Karlsson A, Ingvar M,Ohman A (2004): Fear and the amygdala: Manipulation ofawareness generates differential cerebral responses to phobicand fear-relevant (but nonfeared) stimuli. Emotion 4:340–353.

Compton RJ (2003): The interface between emotion and attention:A review of evidence from psychology and neuroscience.Behav Cogn Neurosci Rev 2:115–129.

Cristinzio C, Sander D, Vuilleumier P (2007): Recognition of emo-tional face expressions and amygdala pathology. Epileptologie24:130–138.

Critchley H, Daly E, Phillips M, Brammer M, Bullmore E, Wil-liams S, Van Amelsvoort T, Robertson D, David A, Murphy D(2000): Explicit and implicit neural mechanisms for processingof social information from facial expressions: A functionalmagnetic resonance imaging study. Hum Brain Mapp 9:93–105.

Critchley HD, Rotshtein P, Nagai Y, O’Doherty J, Mathias CJ,Dolan RJ (2005): Activity in the human brain predicting differ-ential heart rate responses to emotional facial expressions.Neuroimage 24:751–762.

de Gelder B, Vroomen J, Pourtois G, Weiskrantz L (1999): Non-conscious recognition of affect in the absence of striate cortex.Neuroreport 10:3759–3763.

de Gelder B, Morris JS, Dolan RJ (2005): Unconscious fear influen-ces emotional awareness of faces and voices. Proc Natl AcadSci USA 102:18682–18687.

Di Russo F, Martinez A, Sereno MI, Pitzalis S, Hillyard SA (2002):Cortical sources of the early components of the visual evokedpotential. Hum Brain Mapp 15:95–111.

Di Russo F, Martinez A, Hillyard SA (2003): Source analysis ofevent-related cortical activity during visuo-spatial attention.Cereb Cortex 13:486–499.


r 1103 r

Dolan RJ (2002): Emotion, cognition, and behavior. Science298:1191–1194.

Dolan RJ, Vuilleumier P (2003): Amygdala automaticity in emo-tional processing. Ann N Y Acad Sci 985:348–355.

Eger E, Jedynak A, Iwaki T, Skrandies W (2003): Rapid extractionof emotional expression: Evidence from evoked potential fieldsduring brief presentation of face stimuli. Neuropsychologia41:808–817.

Eimer M, Holmes A (2002): An ERP study on the time course ofemotional face processing. Neuroreport 13:427–431.

Eimer M, Holmes A (2007): Event-related brain potential corre-lates of emotional face processing. Neuropsychologia 45:15–31.

Eimer M, Kiss M, Holmes A (2008): Links between rapid ERPresponses to fearful faces and conscious awareness. J Neuro-psychol 2:165–181.

Ekman P, Friesen WV (1976): Pictures of facial affect. Palo Alto,CA: Consulting Psychologist Press.

Friston KJ, Penny WD, Glaser DE (2005): Conjunction revisited.Neuroimage 25:661–667.

George N, Evans J, Fiori N, Davidoff J, Renault B (1996): Brainevents related to normal and moderately scrambled faces.Brain Res Cogn Brain Res 4:65–76.

Gothard KM, Battaglia FP, Erickson CA, Spitler KM, Amaral DG(2007): Neural responses to facial expression and face identityin the monkey amygdala. J Neurophysiol 97:1671–1683.

Graham R, Devinsky O, LaBar KS (2006): Sequential ordering ofmorphed faces and facial expressions following temporal lobedamage. Neuropsychologia 44:1398–1405.

Hamann SB, Adolphs R (1999): Normal recognition of emotionalsimilarity between facial expressions following bilateral amyg-dala damage. Neuropsychologia 37:1135–1141.

Holmes A, Vuilleumier P, Eimer M (2003): The processing of emo-tional facial expression is gated by spatial attention: Evidencefrom event-related brain potentials. Brain Res Cogn Brain Res16:174–184.

Johansson M, Mecklinger A, Treese AC (2004): Recognition mem-ory for emotional and neutral faces: An event-related potentialstudy. J Cogn Neurosci 16:1840–1853.

Keil A, Muller MM, Gruber T, Wienbruch C, Stolarova M, ElbertT (2001): Effects of emotional arousal in the cerebral hemi-spheres: A study of oscillatory brain activity and event-relatedpotentials. Clin Neurophysiol 112:2057–2068.

Kiebel SJ, Friston KJ (2004a) Statistical parametric mapping forevent-related potentials (II): A hierarchical temporal model.Neuroimage 22:503–520.

Kiebel SJ, Friston KJ (2004b) Statistical parametric mapping forevent-related potentials: I. Generic considerations. Neuroimage22:492–502.

Kissler J, Assadollahi R, Herbert C (2006): Emotional and semanticnetworks in visual word processing: Insights from ERP stud-ies. Prog Brain Res 156:147–183.

Krolak-Salmon P, Henaff MA, Vighetto A, Bertrand O, MauguiereF (2004): Early amygdala reaction to fear spreading in occipital,temporal, and frontal cortex: A depth electrode ERP study inhuman. Neuron 42:665–676.

LaBar KS, LeDoux JE, Spencer DD, Phelps EA (1995): Impairedfear conditioning following unilateral temporal lobectomy inhumans. J Neurosci 15:6846–6855.

LeDoux JE (1996): The Emotional Brain. New York: Simon andSchuster.

Leppanen JM, Kauppinen P, Peltola MJ, Hietanen JK (2007): Dif-ferential electrocortical responses to increasing intensities of

fearful and happy emotional expressions. Brain Res 1166:103–109.

Liddell BJ, Williams LM, Rathjen J, Shevrin H, Gordon E (2004): Atemporal dissociation of subliminal versus supraliminal fearperception: An event-related potential study. J Cogn Neurosci16:479–486.

Liddell BJ, Brown KJ, Kemp AH, Barton MJ, Das P, Peduto A,Gordon E, Williams LM (2005): A direct brainstem-amygdala-cortical ‘alarm’ system for subliminal signals of fear. Neuro-image 24:235–243.

Liu J, Harris A, Kanwisher N (2002): Stages of processing in faceperception. Nat Neurosci 5:910–916.

Maratos EJ, Allan K, Rugg MD (2000): Recognition memory foremotionally negative and neutral words: An ERP study. Neu-ropsychologia 38:1452–1465.

Marzi CA, Girelli M, Miniussi C, Smania N, Maravita A (2000):Electrophysiological correlates of conscious vision: Evidencefrom unilateral extinction. J Cogn Neurosci 12:869–877.

McKelvie SJ (1995): Emotional expression in upside-down faces:Evidence for configurational and componential processing. Br JSoc Psychol 34 (Pt 3):325–334.

Meletti S, Benuzzi F, Rubboli G, Cantalupo G, Stanzani MM,Nichelli P, Tassinari CA (2003): Impaired facial emotion recog-nition in early-onset right mesial temporal lobe epilepsy.Neurology 60:426–431.

Mikhailova ES, Davydov DV (1999): Visual evoked potentials inhumans during recognition of emotional facial expressions.Neurosci Behav Physiol 29:687–694.

Morris JS, Friston KJ, Buchel C, Frith CD, Young AW, Calder AJ,Dolan RJ (1998): A neuromodulatory role for the human amyg-dala in processing emotional facial expressions. Brain 121 (Pt1):47–57.

Morris JS, Ohman A, Dolan RJ (1999): A subcortical pathway tothe right amygdala mediating ‘‘unseen’’ fear. Proc Natl AcadSci USA 96:1680–1685.

Morris JS, DeGelder B, Weiskrantz L, Dolan RJ (2001): Differentialextrageniculostriate and amygdala responses to presentation ofemotional faces in a cortically blind field. Brain 124:1241–1252.

Noesselt T, Driver J, Heinze HJ, Dolan R (2005): Asymmetricalactivation in the human brain during processing of fearfulfaces. Curr Biol 15:424–429.

Ohman A (2005): The role of the amygdala in human fear: Auto-matic detection of threat. Psychoneuroendocrinology 30:953–958.

Ohman A, Carlsson K, Lundqvist D, Ingvar M (2007): On theunconscious subcortical origin of human fear. Physiol Behav92:180–185.

Oya H, Kawasaki H, Howard MA, III, Adolphs R (2002): Electro-physiological responses in the human amygdala discriminateemotion categories of complex visual stimuli. J Neurosci 22:9502–9512.

Pegna AJ, Khateb A, Lazeyras F, Seghier ML (2005): Discriminat-ing emotional faces without primary visual cortices involvesthe right amygdala. Nat Neurosci 8:24–25.

Penny W, Holmes A, Friston KJ.2003. Random effects analysis. In:Frackowiak RS, Friston KJ, Frith C, Dolan RJ, Price C, Zeki S,Ashburner J, Penny W. editors. Human Brain Function.Oxford: Academic Press. pp 843–850.

Pessoa L, Kastner S, Ungerleider LG (2002): Attentional control ofthe processing of neural and emotional stimuli. Brain ResCogn Brain Res 15:31–45.


r 1104 r

Pessoa L, Japee S, Sturman D, Ungerleider LG (2005a) Target visi-bility and visual awareness modulate amygdala responses tofearful faces. Cereb Cortex 16:366–375.

Pessoa L, Padmala S, Morland T (2005b) Fate of unattended fear-ful faces in the amygdala is determined by both attentionalresources and cognitive modulation. Neuroimage 28:249–255.

Phelps EA, LeDoux JE (2005): Contributions of the amygdala toemotion processing: From animal models to human behavior.Neuron 48:175–187.

Phillips ML, Williams LM, Heining M, Herba CM, Russell T,Andrew C, Bullmore ET, Brammer MJ, Williams SC, MorganM, Young AW, Gray JA (2004): Differential neural responses toovert and covert presentations of facial expressions of fear anddisgust. Neuroimage 21:1484–1496.

Pourtois G, Grandjean D, Sander D, Vuilleumier P (2004): Electro-physiological correlates of rapid spatial orienting towards fear-ful faces. Cereb Cortex 14:619–633.

Pourtois G, Dan ES, Grandjean D, Sander D, Vuilleumier P(2005a) Enhanced extrastriate visual response to bandpass spa-tial frequency filtered fearful faces: Time course and topo-graphic evoked-potentials mapping. Hum Brain Mapp 26:65–79.

Pourtois G, Thut G, Grave de PR, Michel C, Vuilleumier P (2005b)Two electrophysiological stages of spatial orienting towardsfearful faces: Early temporo-parietal activation preceding gaincontrol in extrastriate visual cortex. Neuroimage 26:149–163.

Rapcsak SZ, Galper SR, Comer JF, Reminger SL, Nielsen L, Kasz-niak AW, Verfaellie M, Laguna JF, Labiner DM, Cohen RA(2000): Fear recognition deficits after focal brain damage: Acautionary note. Neurology 54:575–581.

Richardson MP, Strange BA, Dolan RJ (2004): Encoding of emo-tional memories depends on amygdala and hippocampus andtheir interactions. Nat Neurosci 7:278–285.

Rotshtein P, Malach R, Hadar U, Graif M, Hendler T (2001): Feel-ing or features: Different sensitivity to emotion in high-ordervisual cortex and amygdala. Neuron 32:747–757.

Siebert M, Markowitsch HJ, Bartel P (2003): Amygdala, affect andcognition: Evidence from 10 patients with Urbach–Wiethe dis-ease. Brain 126:2627–2637.

Smith AP, Stephan KE, Rugg MD, Dolan RJ (2006): Task and con-tent modulate amygdala-hippocampal connectivity in emo-tional retrieval. Neuron 49:631–638.

Sprengelmeyer R, Jentzsch I (2006): Event related potentials andthe perception of intensity in facial expressions. Neuropsycho-logia 44:2899–2906.

Streit M, Dammers J, Simsek-Kraues S, Brinkmeyer J, Wolwer W,Ioannides A (2003): Time course of regional brain activationsduring facial emotion recognition in humans. Neurosci Lett342:101–104.

Talmi D, Schimmack U, Paterson T, Moscovitch M (2007): Therole of attention and relatedness in emotionally enhancedmemory. Emotion 7:89–102.

Thorpe S, Fize D, Marlot C (1996): Speed of processing in thehuman visual system. Nature 381:520–522.

van Heijnsbergen CC, Meeren HK, Grezes J, de GB (2007): Rapiddetection of fear in body expressions, an ERP study. Brain Res1186:233–241.

Vuilleumier P (2005): How brains beware: Neural mechanisms ofemotional attention. Trends Cogn Sci 9:585–594.

Vuilleumier P, Armony JL, Driver J, Dolan RJ (2001): Effects ofattention and emotion on face processing in the human brain:An event-related fMRI study. Neuron 30:829–841.

Vuilleumier P, Richardson MP, Armony JL, Driver J, Dolan RJ(2004): Distant influences of amygdala lesion on visual corticalactivation during emotional face processing. Nat Neurosci7:1271–1278.

Williams LM, Brown KJ, Das P, Boucsein W, Sokolov EN,Brammer MJ, Olivieri G, Peduto A, Gordon E (2004): The dy-namics of cortico-amygdala and autonomic activity over theexperimental time course of fear perception. Brain Res CognBrain Res 21:114–123.


r 1105 r

Documents

Amygdala damage affects event-related potentials for fearful faces at specific time windows